in reply to Re^2: POE and recursion
in thread POE and recursion

This kind of recursion can be rewritten as iteration. Here's some untested code:

my @subcategories = ("Physics"); my %unique_pages; while (@subcategories) { my $next_subcategory = shift @subcategories; my @nodes = fetch_subcategory($next_subcategory); NODE: foreach my $node (@nodes) { if ($node->type() eq "subcategory") { push @subcategories, $node; next NODE; } $unique_pages{$node->unique_id()} = $node; } } foreach my $page (values %unique_pages) { my $page_content = fetch_page($page); # ... stuff here ... }

In POE, you might keep the subcategories list and the unique pages hash in a session's heap. Class-scoped lexicals are also good here, especially if you plan for many sessions to share the burden of crawling. Here's one way to do it:

sub fetch_next_subcategory { my ($kernel, $heap) = @_[KERNEL, HEAP]; my $next_subcategory = shift @{$heap->{subcategories}}; if (defined $next_subcategory) { # ... start a subcategory fetch ... } else { # switch to fetching articles $kernel->yield("fetch_next_article"); } } sub got_a_subcategory_page { my ($kernel, $heap, $page) = @_[KERNEL, HEAP, ARG0]; my @nodes = parse_subcategory($page); NODE: foreach my $node (@nodes) { if ($node->type() eq "subcategory") { push @{$heap->{subcategories}}, $node; } else { push @{$heap->{pages}}, $node; } } $kernel->yield("fetch_next_subcategory"); } sub fetch_next_article { my ($kernel, $heap) = @_[KERNEL, HEAP]; my $next_page; while (defined($next_page = shift @{$heap->{pages}})) { last unless exists $heap->{unique_pages}{$next_page->unique_id()}; } unless (defined $next_page) { # ... done. return; } $heap->{unique_pages}{$next_page->unique_id()} = 1; # ... start fetching next page } sub got_an_article_page { my ($kernel, $heap) = @_[KERNEL, HEAP]; # ... process page. $kernel->yield("fetch_next_article"); }